A method for improving data transmission of self-learning intelligent scale

Abstract

The present invention discloses a method for improving data transmission of a self-learning intelligent scale, comprising: a plurality of main scales start self-learning, record and store self-learning data in a local memory of the corresponding main scale; the self-learning main scale compresses the self-learning data packet; the self-learning main scale sends the compressed self-learning data packet to a communication channel, transmits the self-learning data between the self-learning main scales, and aggregates the self-learning data; the self-learning main scale sends the aggregated self-learning data packet to a central system, and the aggregated self-learning data packet is used to trigger the central system to distribute the aggregated self-learning data packet to the main scale and all the auxiliary scales that have not performed self-learning; wherein all the main scales, auxiliary scales and the central system apply error detection and error correction mechanisms, and the communication channel is planned by time division multiplexing and frequency division multiplexing methods. The present invention optimizes the communication channel by combining time division multiplexing and frequency division multiplexing methods, and realizes efficient data transmission between the intelligent scale and the central system.

Description

Technical Field

[0001] The present invention relates to the technical field of digital information transmission, in particular to the technical field combined with the use of different frequencies, and specifically to a method for improving data transmission of a self-learning intelligent scale. Background Art

[0002] With the rapid development of the retail industry such as fresh food and snacks, stores have an increasing demand for efficient and accurate product identification and weighing. Traditional weighing methods often rely on manual operations, which are not only inefficient but also prone to errors, resulting in inaccurate data. To solve this problem, self-learning smart scales have emerged. This smart scale can automatically identify goods, measure weight, and record relevant data through built-in sensors and algorithms, thereby significantly improving cash register efficiency. However, with the expansion of store size and the increase in product types, the efficiency of data transmission between smart scales and central systems has become an urgent problem to be solved. In order to improve data transmission efficiency and ensure the real-time, accuracy and security of data, it is necessary to study and apply a series of advanced data transmission technologies and strategies.

[0003] Although self-learning smart scales have greatly improved the efficiency of commodity identification and weighing, the existing technology still has some obvious defects in data transmission: 1. The data packet structure is often not compact enough and contains a lot of redundant information, resulting in excessive data transmission and increased bandwidth requirements; 2. There is a lack of effective error control and correction mechanism, and data is prone to errors or loss during transmission, affecting the integrity and accuracy of the data; 3. The application of multi-channel communication technology is insufficient, and communication resources cannot be fully utilized to achieve simultaneous communication of multiple smart scales; 4. There is a lack of effective performance monitoring and optimization means, and it is impossible to monitor the data transmission status in real time and adjust the transmission parameters according to the monitoring results to optimize the data transmission efficiency. These defects limit the application and promotion of self-learning smart scales in larger-scale stores.

[0004] Therefore, it is necessary to improve the data transmission improvement method of the self-learning smart scale in the prior art to solve the above problems. Summary of the invention

[0005] The present invention overcomes the shortcomings of the prior art and provides a method for improving data transmission of a self-learning intelligent scale, aiming to solve the problems in the prior art of self-learning intelligent scales in data transmission, such as a non-compact data packet structure, a large amount of redundant information, a lack of effective error control and correction mechanism, insufficient application of multi-channel communication technology, and a lack of effective performance monitoring and optimization means.

[0006] To achieve the above object, the technical solution adopted by the present invention is: a data transmission improvement method for a self-learning intelligent scale, comprising:

[0007] Several main scales start self-learning, and record and store the self-learning data in the local memory of the corresponding main scales;

[0008] The self-learning master scale compresses the self-learning data packet;

[0009] The self-learning master scale sends the compressed self-learning data packet to the communication channel, transmits the self-learning data between the self-learning master scales, and aggregates the self-learning data;

[0010] The self-learning master scale sends the aggregated self-learning data packets to the central system, and the aggregated self-learning data packets are used to trigger the central system to distribute the aggregated self-learning data packets to the master scale and all slave scales that have not performed self-learning;

[0011] Among them, all main scales, auxiliary scales and central systems apply error detection and error correction mechanisms, and communication channels are planned through time division multiplexing and frequency division multiplexing methods.

[0012] In a preferred embodiment of the present invention, the steps of collecting and storing self-learning data include:

[0013] The main scale starts self-learning, and uses its built-in sensors and algorithms to identify goods and measure weight, and records the main scale self-learning data in its local storage;

[0014] The central system detects that the main scale starts self-learning, and issues data collection instructions to several secondary scales. The secondary scales communicate with the corresponding main scales and upload the commodity-related data of the secondary scales to the corresponding connected main scales;

[0015] After receiving the secondary scale commodity-related data uploaded by the secondary scale, the main scale integrates it with the main scale's self-learning data and stores it in its local storage.

[0016] In a preferred embodiment of the present invention, the steps of the error detection mechanism include:

[0017] At the port sending the self-learning data, a cyclic redundancy check algorithm is used to generate a first check code for the compressed self-learning data packet, and the first check code is added to the end of the self-learning data packet;

[0018] At the port receiving the self-learning data, a cyclic redundancy check algorithm is used to recalculate the second check code of the received self-learning data packet, and the second check code is compared with the first check code;

[0019] When the second check code matches the first check code, there is no problem with the transmission; when the second check code does not match the first check code, an error occurs during the transmission of the self-learning data packet.

[0020] In a preferred embodiment of the present invention, when the error detection mechanism detects that an error occurs in the transmission process of the self-learning data packet, the error correction mechanism is triggered;

[0021] At the port sending the self-learning data, Reed-Solomon coding is used to add redundant information to the self-learning data packet;

[0022] When the second check code does not match the first check code, a Reed-Solomon decoder is used at the port receiving the self-learning data to correct data transmission errors in the self-learning data packet.

[0023] In a preferred embodiment of the present invention, the codeword length of the Reed-Solomon code is Sum check symbol digits , , is the number of self-learning data bits after splicing; The number of errors corrected and detected for Reed-Solomon codes; , The number of errors to be corrected;

[0024] In a finite field Generating polynomials ,in, It is the original element. is a formal variable used to represent the terms of the polynomial; suppose the concatenated self-learning dataset , the corresponding polynomial , the check symbol polynomial is , calculate the corresponding check symbol of the self-learning data, and splice the check symbol into the self-learning data set to form the encrypted self-learning data code.

[0025] In a preferred embodiment of the present invention, the communication channel planning step includes:

[0026] Determine the total bandwidth of the communication channel and divide it into a number of sub-bands, each of which is used to transmit one signal to achieve frequency division multiplexing; the number of sub-bands must be greater than the number of communication channels, and each sub-band is allocated to at most one communication channel;

[0027] A complete communication cycle is determined, and in each sub-band, the communication cycle is divided into several time slots, each of which is allocated to the main scale and the auxiliary scale for data transmission, thus realizing time division multiplexing.

[0028] In a preferred embodiment of the present invention, a time slot interval is set between time slots, and the time slot interval is between 20-60 microseconds.

[0029] In a preferred embodiment of the present invention, the self-learning data transmission steps between the self-learning master scales are as follows:

[0030] Each self-learning master scale adds the compressed self-learning data packet to the corresponding sub-band in its allocated time slot and sends it to the communication channel;

[0031] The self-learning master scale performs error detection when receiving the self-learning data packet from the communication channel;

[0032] When the data is detected to be correct, the self-learning master scale will collate and verify the received self-learning data packet with its own self-learning data packet to form an aggregated self-learning data packet;

[0033] The aggregated self-learning data packet is updated to the local memory of the self-learning master scale.

[0034] In a preferred embodiment of the present invention, the self-learning data packet can also be parsed, segmented and transmitted, and the specific steps include:

[0035] Before the self-learning master scale sends the self-learning data packet, it divides the self-learning data packet according to the product ID; for each product ID, the self-learning data packet contains all the pictures, weights and timestamps corresponding to the product ID, and is sorted in the order of timestamps;

[0036] A header is added to each segmented self-learning data packet, and the header information includes a serial number, a product ID, and the number of pictures;

[0037] In different time slots, the self-learning master scale sends the segmented data packets to the communication channel;

[0038] When other self-learning master scales receive the split self-learning data packets, they sort them according to the attached serial numbers until all relevant serial numbers have been received, and then perform error detection;

[0039] After error detection, parse the header information, check and update the data in local storage according to the product ID and the number of pictures;

[0040] If data with the same product ID already exists in the local storage, with the same number of images and consistent image content, there is no need to update the data for the product ID;

[0041] Otherwise, only the missing self-learning data will be updated;

[0042] The updated self-learning data is aggregated into an aggregated self-learning data packet.

[0043] In a preferred embodiment of the present invention, the steps of the central system updating the aggregated self-learning data packet are as follows:

[0044] The self-learning master scale sends the aggregated self-learning data packets to the central system, which performs error detection;

[0045] When the data transmission of the aggregated self-learning data packet is correct, the central system distributes the aggregated self-learning data packet to the main scale and all the secondary scales that have not performed self-learning;

[0046] After the main scale and all the auxiliary scales that have not performed self-learning receive the self-learning data from the central system, they update the self-learning data in the local memory.

[0047] The present invention solves the defects existing in the background technology and has the following beneficial effects:

[0048] (1) The present invention collects and stores self-learning data through the main scale and the auxiliary scale, compresses the self-learning data packet to reduce the amount of transmitted data, applies error detection and correction mechanisms to all main scales, auxiliary scales and the central system to ensure the integrity and accuracy of the data, optimizes the communication channel by combining time division multiplexing and frequency division multiplexing methods, and realizes efficient data transmission between multiple smart scales and the central system at the same time. Finally, the main scale sends the aggregated self-learning data packet to the central system, and the central system distributes the data to all main scales and auxiliary scales to realize real-time data update.

[0049] (2) The present invention directly optimizes the data transmission process through the combination of time division multiplexing and frequency division multiplexing, as well as the segmented transmission of self-learning data packets, and can allocate network resources more flexibly. Different small packets can be transmitted in different time slots, making full use of the network's time and space resources. The self-learning data packet is segmented according to the product ID before sending, and each segmented data packet contains relevant data for a specific product ID. This segmentation method reduces the amount of data transmitted in a single transmission, avoids transmission delays or failures caused by excessively large data packets, and further improves the efficiency of data transmission; the receiving end can process multiple small packets at the same time without waiting for the entire large data packet to be fully received before processing, reducing the amount of calculation for data merging, and only needs to update the missing data without parsing all the data, avoiding unnecessary repeated transmission and storage of data.

[0050] (3) The present invention combines time division multiplexing and frequency division multiplexing technologies. Through frequency division multiplexing technology, the total bandwidth of the communication channel is divided into multiple sub-bands. Each sub-band independently transmits one signal, avoiding mutual interference between signals. Using time division multiplexing technology, the communication cycle is divided into multiple time slots in each sub-band and allocated to different main scales and auxiliary scales for data transmission, thereby realizing simultaneous communication of multiple smart scales. The combination of time division and frequency division multiplexing technologies maximizes the use of communication resources, improves the utilization rate of communication channels, and makes data transmission more efficient and orderly.

[0051] (4) The present invention aggregates data between master scales and distributes it to the central system. Each master scale sends a compressed self-learning data packet within the allocated time slot, and performs error detection and data aggregation between the master scales. By exchanging data between the master scales, data verification and validation are performed, so that data errors and inconsistencies can be corrected before being synchronized to the slave scales. Then, the master scale sends the aggregated self-learning data packet to the central system, which distributes it to all master scales and slave scales after verification, thereby realizing real-time data update and synchronization. Data aggregation between master scales reduces the transmission of redundant data and improves data transmission efficiency and accuracy. The unified distribution of the central system ensures the consistency and real-time nature of data from all smart scales, providing accurate and timely data support for store management.

[0052] (5) The present invention integrates error detection and error correction mechanisms, uses a cyclic redundancy check algorithm to generate a checksum, performs error detection on transmitted data packets, and promptly discovers and identifies erroneous data packets; when an error is detected, Reed-Solomon coding is used to add redundant information to the data packet, and the error is corrected at the receiving end by a Reed-Solomon decoder without the need to resend data, thereby improving the reliability of data transmission. The dual protection of the error detection and correction mechanism ensures high integrity and accuracy during the data transmission process, avoiding system errors or decision-making errors caused by data transmission errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.

[0054] Figure 1 is a flow chart of a preferred embodiment of the present invention;

[0055] Figure 2 It is a structural schematic diagram of a preferred embodiment of the present invention;

[0056] Figure 3 is a diagram of an error detection mechanism of a preferred embodiment of the present invention;

[0057] Figure 4 It is a diagram of the error correction mechanism of the preferred embodiment of the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0060] Application Overview:

[0061] There are two types of communication channel planning methods: time division multiplexing and frequency division multiplexing;

[0062] Time division multiplexing is a technology that divides time into multiple time slots, each of which is assigned to a different signal source for transmission. In time division multiplexing, the time domain is divided into a number of small segments that are cyclical, each of which is of fixed length and is used to transmit a subchannel. This improves channel utilization, allowing multiple digital signals to occupy the communication line in time for transmission. However, if the data transmission rate of all signal sources is close to the maximum transmission rate of the channel, the efficiency of time division multiplexing may be reduced.

[0063] Frequency division multiplexing is a technology that divides the total bandwidth of a communication channel into multiple sub-bands, with each sub-band transmitting one signal. The center frequencies of each sub-band do not overlap, and a certain width of isolation bandwidth is reserved between sub-channels to prevent mutual interference between signals. Each sub-channel works in parallel, and there is no need to consider transmission delay when transmitting signals. However, more spectrum resources need to be occupied, and the equipment cost increases with the number of input channels. It is not easy to miniaturize the smart scale suitable for this application.

[0064] In the self-learning smart scale data transmission improvement method of the present application, by combining the use of time division multiplexing and frequency division multiplexing technology, simultaneous data transmission between multiple main scales and efficient data transmission with the central system can be achieved, ensuring the timeliness and reliability of data transmission.

[0065] like Figure 1 and Figure 2 As shown, a method for improving data transmission of a self-learning smart scale includes the following steps:

[0066] Several main scales start self-learning, and record and store the self-learning data in the local memory of the corresponding main scales;

[0067] The self-learning master scale compresses the self-learning data packet;

[0068] The self-learning master scale sends the compressed self-learning data packet to the communication channel, transmits the self-learning data between the self-learning master scales, and aggregates the self-learning data;

[0069] The self-learning master scale sends the aggregated self-learning data packets to the central system, and the aggregated self-learning data packets are used to trigger the central system to distribute the aggregated self-learning data packets to the master scale and all slave scales that have not performed self-learning;

[0070] Among them, all main scales, auxiliary scales and central systems apply error detection and error correction mechanisms, and communication channels are planned through time division multiplexing and frequency division multiplexing methods.

[0071] The main scale is an intelligent scale with more advanced functions or plays a leading role in the network. The secondary scale is an auxiliary scale in the network. It does not have all the advanced functions of the main scale and is used to perform basic commodity identification and weighing tasks.

[0072] The main scale, auxiliary scale and central system together constitute a distributed weighing system. The main scale assumes more coordination and management responsibilities, while the auxiliary scale is mainly responsible for performing basic tasks. Through the distributed weighing system, the efficiency of data transmission and the overall performance of the system can be effectively improved.

[0073] Specifically, the self-learning data includes: product recognition results, weight, timestamp, product images, and each product is assigned a product ID;

[0074] Specifically, the main scale identifies goods and measures weight through built-in sensors and algorithms. The built-in sensors include: cameras, weight sensors, combined with algorithms to identify goods; the algorithm is specifically: preprocessing the original image captured by the camera, using the convolutional neural network deep learning model to extract product features from the preprocessed image. The product features can represent the key information in the product image, and then the target detection algorithm is applied to identify the product in the image.

[0075] Specifically, the steps of collecting and storing self-learning data include:

[0076] The main scale starts self-learning, and uses its built-in sensors and algorithms to identify goods and measure weight, and records the main scale self-learning data in its local storage;

[0077] The central system detects that the main scale starts self-learning, and issues data collection instructions to several secondary scales. The secondary scales communicate with the corresponding main scales and upload the commodity-related data of the secondary scales to the corresponding connected main scales;

[0078] After receiving the secondary scale commodity-related data uploaded by the secondary scale, the main scale integrates it with the main scale's self-learning data and stores it in its local storage.

[0079] Specifically, the integrated self-learning data is organized into a self-learning data packet, and the format of the self-learning data packet is: product image, product ID, weight and timestamp; only key data elements are retained in the self-learning data packet to reduce redundant information;

[0080] Specifically, efficient compression algorithms, including ZIP and GZIP, are used to compress self-learning data packets to reduce the amount of transmitted data, reduce network congestion, and reduce bandwidth requirements.

[0081] Error detection and error correction are important technologies used in communication systems to ensure the integrity and accuracy of data during transmission. During data transmission, data may be erroneous or lost due to signal interference, equipment failure, transmission medium problems, etc. In order to detect and correct these errors, communication systems need to adopt error control and correction mechanisms.

[0082] Error detection refers to checking whether the received self-learning data contains errors in some way; error correction refers to automatically correcting the errors after detecting the errors in the self-learning data without requesting retransmission of the data.

[0083] like Figure 3 As shown, specifically, the steps of the error detection mechanism include:

[0084] At the port sending the self-learning data, a cyclic redundancy check algorithm is used to generate a first check code for the compressed self-learning data packet, and the first check code is added to the end of the self-learning data packet;

[0085] At the port receiving the self-learning data, a cyclic redundancy check algorithm is used to recalculate the second check code of the received self-learning data packet, and the second check code is compared with the first check code;

[0086] When the second check code matches the first check code, there is no problem with the transmission; when the second check code does not match the first check code, an error occurs during the transmission of the self-learning data packet;

[0087] like Figure 4 As shown, specifically, when the error detection mechanism detects that an error occurs in the transmission process of the self-learning data packet, the error correction mechanism is triggered;

[0088] At the port sending the self-learning data, Reed-Solomon coding is used to add redundant information to the self-learning data packet;

[0089] When the second check code does not match the first check code, a Reed-Solomon decoder is used at the port receiving the self-learning data to correct data transmission errors in the self-learning data packet.

[0090] Reed-Solomon coding is an error correction coding technique widely used in digital communication and storage systems. It enhances the data's error resistance by adding redundant data.

[0091] Determine the codeword length of Reed-Solomon codes Sum check symbol digits , , is the number of self-learning data bits after splicing; is the number of errors that can be corrected and detected by the Reed-Solomon code; , is the number of errors that can be corrected;

[0092] In a finite field Generating polynomials ,in, It is the original element. is a formal variable used to represent the terms of the polynomial; suppose the concatenated self-learning dataset , the corresponding polynomial , the check symbol polynomial is , calculate the corresponding check symbol of the self-learning data, and splice the check symbol into the self-learning data set to form the encrypted self-learning data code.

[0093] In the data transmission between the smart scale and the central system, the application of error control and correction mechanisms can significantly improve the reliability of data transmission. By detecting and correcting errors in the transmission process, it can ensure that the received data is complete and accurate, thereby avoiding system failures or wrong decisions caused by data errors.

[0094] Time division multiplexing is a technology that divides time into multiple time slots and assigns each time slot to a different signal source for transmission. Frequency division multiplexing is a technology that divides the total bandwidth of a communication channel into multiple sub-bands, each of which transmits one signal. Combining the two enables multiple smart scales to simultaneously transmit data efficiently with the central system, ensuring the timeliness and reliability of data transmission.

[0095] Specifically, the planning steps of the communication channel include:

[0096] Determine the total bandwidth of the communication channel and divide it into several sub-bands, each of which is used to transmit one signal to achieve frequency division multiplexing;

[0097] The number of sub-bands should be greater than the number of communication channels. Each sub-band is assigned to at most one communication channel to ensure that the sub-bands of the communication channels do not overlap with each other to avoid interference. The sub-bands of the communication channels for information transmission between smart scales are allocated according to the type of smart scales, the flow of people in the area where the smart scales are located, and the data transmission requirements. Assume that the total bandwidth of the system is Hz, the total bandwidth Split into sub-bands, the width of each sub-band ;

[0098] As the flow of people in an area changes over time, the allocation of redundant sub-bands is flexible, and sub-bands can be freely reorganized and split to achieve dynamic allocation of bands for easy adjustment.

[0099] Determine a complete communication cycle, divide the communication cycle into several time slots in each sub-band, and allocate each time slot to the main scale and the auxiliary scale for data transmission to realize time division multiplexing;

[0100] According to the data volume and data transmission requirements of the smart scale including the main scale and the auxiliary scale, the length of the corresponding time slot is set to ensure that each scale has enough time to send its data; set a complete communication cycle , in each sub-band, the communication cycle is divided into time slots;

[0101] Specifically, in the time slot Set the time slot interval between channels to avoid conflicts on the channel. Between 20-60 microseconds; the time slot interval and the time slot constitute a complete communication cycle ;

[0102] Specifically, the steps for self-learning data transmission between self-learning master scales are as follows:

[0103] Each self-learning master scale adds the compressed self-learning data packet to the corresponding sub-band in its allocated time slot and sends it to the communication channel;

[0104] The self-learning master scale performs error detection when receiving the self-learning data packet from the communication channel;

[0105] When the data is detected to be correct, the self-learning master scale will collate and verify the received self-learning data packet with its own self-learning data packet to form an aggregated self-learning data packet;

[0106] The aggregated self-learning data packets are updated to the local memory of the self-learning master scale;

[0107] However, the self-learning data packet contains a large amount of data, such as product images, recognition results, weight, and timestamps. If the data packet is too large, direct transmission may put pressure on the network bandwidth and reduce transmission efficiency. The combination of time division multiplexing and frequency division multiplexing alone cannot directly solve this problem. In this case:

[0108] Furthermore, the self-learning data packet is parsed and segmented for transmission. The specific steps include:

[0109] Before the self-learning master scale sends the self-learning data packet, it divides the self-learning data packet according to the product ID; for each product ID, the self-learning data packet contains all the pictures, weights and timestamps corresponding to the product ID, and is sorted in the order of timestamps;

[0110] A header is added to each segmented self-learning data packet, and the header information includes a serial number, a product ID, and the number of pictures;

[0111] Sequence numbers ensure that the receiving end can reassemble packets in the correct order; self-learning packet segmentation helps prevent transmission failures due to overly large packets, as well as avoid packet sticking and unpacking problems caused by network delays or packet loss;

[0112] In different time slots, the self-learning master scale sends the segmented data packets to the communication channel; when other self-learning master scales receive the segmented self-learning data packets, they sort them according to the attached sequence numbers until all relevant sequence numbers are received, and then perform error detection;

[0113] After error detection, parse the header information, check and update the data in local storage according to the product ID and the number of pictures;

[0114] If data with the same product ID already exists in the local storage, with the same number of images and consistent image content, there is no need to update the data for the product ID;

[0115] Otherwise, only the missing self-learning data will be updated;

[0116] The updated self-learning data is aggregated into an aggregated self-learning data packet.

[0117] Splitting the self-learning data packet allows the receiver to process multiple small packets at the same time without having to wait for the entire large data packet to be fully received before processing, reducing the amount of computational effort for data merging. It only needs to update the missing data for a certain product without having to parse and update all the data, thus avoiding unnecessary repeated transmission and storage of data.

[0118] Clock synchronization between smart scales and central systems uses synchronization signals to control the start and end of data transmission.

[0119] The self-learning data is aggregated between the self-learning master scales first, which reduces data redundancy and helps ensure the consistency and accuracy of data between different master scales.

[0120] By self-learning the data exchange between the main scales, data verification and validation can be performed to detect potential errors or inconsistencies and correct them before synchronizing to the secondary scales.

[0121] Specifically, the steps for the central system to update the aggregated self-learning data package are as follows:

[0122] The self-learning master scale sends the aggregated self-learning data packets to the central system, which performs error detection;

[0123] When the data transmission of the aggregated self-learning data packet is correct, the central system distributes the aggregated self-learning data packet to the main scale and all the secondary scales that have not performed self-learning;

[0124] After the main scale and all the auxiliary scales that have not performed self-learning receive the self-learning data from the central system, they update the self-learning data in the local memory.

[0125] The above is based on the ideal embodiment of the present invention. Through the above description, relevant personnel can make various changes and modifications without departing from the technical concept of the present invention. The technical scope of the present invention is not limited to the content in the specification, and the technical scope must be determined according to the scope of the claims.

Claims

1. A method for improving data transmission of a self-learning intelligent scale, characterized in that: Includes steps: Several main scales start self-learning, and record and store the self-learning data in the local memory of the corresponding main scales; The self-learning master scale compresses the self-learning data packet; The self-learning master scale sends the compressed self-learning data packet to the communication channel, transmits the self-learning data between the self-learning master scales, and aggregates the self-learning data; The self-learning master scale sends the aggregated self-learning data packets to the central system, and the aggregated self-learning data packets are used to trigger the central system to distribute the aggregated self-learning data packets to the master scale and all slave scales that have not performed self-learning; Among them, all main scales, auxiliary scales and central systems apply error detection and error correction mechanisms, and communication channels are planned through time division multiplexing and frequency division multiplexing methods; The steps of the error detection mechanism include: At the port sending the self-learning data, a cyclic redundancy check algorithm is used to generate a first check code for the compressed self-learning data packet, and the first check code is added to the end of the self-learning data packet; At the port receiving the self-learning data, a cyclic redundancy check algorithm is used to recalculate the second check code of the received self-learning data packet, and the second check code is compared with the first check code; When the second check code matches the first check code, there is no problem with the transmission; when the second check code does not match the first check code, an error occurs during the transmission of the self-learning data packet; When the error detection mechanism detects that an error occurs in the transmission process of the self-learning data packet, the error correction mechanism is triggered; At the port sending the self-learning data, Reed-Solomon coding is used to add redundant information to the self-learning data packet; When the second check code does not match the first check code, a Reed-Solomon decoder is used at the port receiving the self-learning data to correct a data transmission error in the self-learning data packet; The steps for planning a communication channel include: Determine the total bandwidth of the communication channel and divide it into a number of sub-bands, each of which is used to transmit one signal to achieve frequency division multiplexing; the number of sub-bands must be greater than the number of communication channels, and each sub-band is allocated to at most one communication channel; A complete communication cycle is determined, and in each sub-band, the communication cycle is divided into several time slots, each of which is allocated to the main scale and the auxiliary scale for data transmission, thus realizing time division multiplexing.

2. The method for improving data transmission of a self-learning intelligent scale according to claim 1, characterized in that: The steps for collecting and storing self-learning data include: The main scale starts self-learning, and uses its built-in sensors and algorithms to identify goods and measure weight, and records the main scale self-learning data in its local storage; The central system detects that the main scale starts self-learning, and issues data collection instructions to several secondary scales. The secondary scales communicate with the corresponding main scales and upload the commodity-related data of the secondary scales to the corresponding connected main scales; After receiving the secondary scale commodity-related data uploaded by the secondary scale, the primary scale integrates it with the primary scale's self-learning data and stores it in its local storage.

3. The method for improving data transmission of a self-learning intelligent scale according to claim 1, characterized in that: Assume the codeword length n of the Reed-Solomon code and the number of check symbol bits r, n = k + r, k is the number of self-learning data bits after splicing; d is the number of errors corrected and detected by the Reed-Solomon code; d = 2m + 1, m is the number of errors corrected; In a finite field GF(2 r ) Generating polynomial g(x) = (x-1)(x-α)…(x-α r-1 ), where α is a primitive element and x is a formal variable used to represent the terms of the polynomial; let the concatenated self-learning data set S = (s0, s1, …, s k-1 ), the corresponding polynomial S(x) = s0+s1x+…+s k-1 x k-1 , the check symbol polynomial is R(x)=S(x)·g(x)modx n , calculate the corresponding check symbol of the self-learning data, and splice the check symbol into the self-learning data set to form the encrypted self-learning data code.

4. The method for improving data transmission of a self-learning intelligent scale according to claim 1, characterized in that: The time slot interval is set between the time slots, and the time slot interval is between 20-60 microseconds.

5. The method for improving data transmission of a self-learning intelligent scale according to claim 1, characterized in that: Self-learning data transmission steps between master scales: Each self-learning master scale adds the compressed self-learning data packet to the corresponding sub-band in its allocated time slot and sends it to the communication channel; The self-learning master scale performs error detection when receiving the self-learning data packet from the communication channel; When the data is detected to be correct, the self-learning master scale will collate and verify the received self-learning data packet with its own self-learning data packet to form an aggregated self-learning data packet; The aggregated self-learning data packet is updated to the local memory of the self-learning master scale.

6. The method for improving data transmission of a self-learning intelligent scale according to claim 5, characterized in that: The self-learning data packet can also be parsed and segmented for transmission. The specific steps include: Before the self-learning master scale sends the self-learning data packet, it divides the self-learning data packet according to the product ID; for each product ID, the self-learning data packet contains all the pictures, weights and timestamps corresponding to the product ID, and is sorted in the order of timestamps; A header is added to each segmented self-learning data packet, and the header information includes a serial number, a product ID, and the number of pictures; In different time slots, the self-learning master scale sends the segmented data packets to the communication channel; When other self-learning master scales receive the split self-learning data packets, they sort them according to the attached serial numbers until all relevant serial numbers have been received, and then perform error detection; After error detection, parse the header information, check and update the data in local storage according to the product ID and the number of pictures; If data with the same product ID already exists in the local storage, with the same number of images and consistent image content, there is no need to update the data for the product ID; Otherwise, only the missing self-learning data will be updated; The updated self-learning data is aggregated into an aggregated self-learning data packet.

7. The method for improving data transmission of a self-learning intelligent scale according to claim 1, characterized in that: Steps for the central system to update the aggregated self-learning data package: The self-learning master scale sends the aggregated self-learning data packets to the central system, which performs error detection; When the data transmission of the aggregated self-learning data packet is correct, the central system distributes the aggregated self-learning data packet to the main scale and all the secondary scales that have not performed self-learning; After the main scale and all the auxiliary scales that have not performed self-learning receive the self-learning data from the central system, they update the self-learning data in the local memory.

Images